← Back to team overview

canonical-ci-engineering team mailing list archive

Obscure autopilot tracebacks, we can help them help us ; )

 

I had a discussion with thomi during the sprint about some tracebacks
from autopilot tests that I found hard to read and he agreed to work on
improving them.

Since then I've forwarded some I ran into and he was very receptive,
file bugs for them and even fixed some of those bugs already (well, one,
in ap-1.4, but hey, it's fixed ;). This includes http://pad.lv/1237035 ,
http://pad.lv/1237039 and http://pad.lv/1239427

http://pad.lv/1237022 is also worth mentioning as it's the culprit for a
lot of noise in the jenkins logs like:

unknown option -- t
unknown option -- e
unknown option -- s
unknown option -- t
unknown option -- a
unknown option -- b
unknown option -- i
unknown option -- l
unknown option -- i
unknown option -- t
unknown option -- y

more people subscribing to any of these bugs may help get them fixed
sooner ;)

I think it's an important way to improve the whole ci engine S/N ratio
to have the logs and test results contains better indications of what
went wrong for the developer discovering a test failure.

We're bound to be confronted more and more with people seeing test
failures caused by other projects and I personally am already tired to
have to filter so much noise ;)

     Vincent